How many nodes can leave the network until data loss may occur? Are there any statistics / dashboards for this question?
Edit: If you think this is too offtopic i can also start a new thread…
Whatever the change may be, I imagine it would be implemented gradually to prevent possible churn from happening at the same time. In that scenario, repair can take care of any nodes that leave. Hopefully SNOs who do decide to leave will run graceful exit to get their held amount too. The network can handle a lot.
Also, the first step will just be a proposal, I believe. So if the feedback on that is really bad, I’m guessing they will adapt.
2 Likes
Whatever the change may be, I imagine it would be implemented gradually to prevent possible churn from happening at the same time. In that scenario, repair can take care of any nodes that leave. Hopefully SNOs who do decide to leave will run graceful exit to get their held amount too.
- I have some nodes with held amount > 50$ => a graceful exit would be a good idea.
- Other nodes have only 1-2 $ - so gracefully exiting wouldn’t cover energy costs => hard shutdown?!
The network can handle a lot.
a lot is not very precise. any hard facts?
Also, the first step will just be a proposal, I believe. So if the feedback on that is really bad, I’m guessing they will adapt.
The feedback WILL be very bad. Otherwise they would just released the paper/change the payout rates. I personally think that there won’t be much room to negotiate. So everyone get their popcorn ready 
2 Likes
Let’s not get ahead of ourselves and see what the proposal is. If you do want to exit, you’ll have to make the determination yourself. I would suggest running graceful exit in everything, but I guess people can’t be expected to take a loss for the good of the network.
As for solid numbers, you can calculate how much Storj could instantly lose. But it would be a quite meaningless numbers as nodes won’t go offline in a coordinated fashion like that. It depends much more on how fast they can repair data if needed and I don’t think those numbers are publicly available. However I do believe they have the ability to scale up repair workers if required.
Keep in mind that you need only 29 pieces out of 80 to be able to get the file. I believe repair currently kicks in at around 54 pieces, so that would be the lowest availability on the network. So there is a lot of slack built in in case there is churn.
As for negotiating… We don’t need to. If there is a high risk of high churn, that risk does the negotiating for us as Storj can’t risk file loss or even losing too much capacity on the network. So I’m actually quite certain our feedback matters. If the new system doesn’t work for you, let them know why and tell them if you plan to exit. That way they will have to adapt if too many people respond like that.
But for now, let’s wait for more news first.
7 Likes
I think it will be a difficult thing to try and balance out. No matter what number is ultimately decided on, there are going to be people posting their configuration, power usage, costs per kwh, and how they can’t earn anything with whatever the numbers are. We have that now with some people. We can make recommendations to run low wattage devices and on hardware that is already running for other purposes, but each person is going to have to decide what their breaking point is for earning or not.
One thing to keep in mind though, is that those who leave, their data will likely be added to those who stay, and so the ones who stay may balance out by having a lot more data than they had before. Whether that works against the new economic model is going to be known soon enough.
5 Likes
It comes at a bad time. Energy prices are at an all time high for many people atm. For me that is going to get slightly better in April and I’m looking to switch to a cheaper power company as well. But if this would have happened with the prices I had early 2022, I could have handled a much bigger drop in payouts. That said, my biggest nodes aren’t going anywhere as they run on always on hardware anyway. It’s just the external USB connected nodes that I may have to reconsider. Which at the moment is less than a third of my total stored data. And depending on how things go, I might actually migrate those to my internal array. We’ll see.
The proposal has now been posted. Let’s lock this topic and continue the discussion there.
Well, the proposal is out https://forum.storj.io/t/update-proposal-for-storage-node-operators/21749
Basically the same thing that OP proposed ($1 per TB stored
$5 per TB egress (including repair/audit) and not a word about changing price to customers
1 Like